Zebrafish drug screening identifies candidate therapies for neuroprotection after spontaneous intracerebral haemorrhage

Abstract

Despite the global health burden, treatment of spontaneous intracerebral haemorrhage (ICH) is largely supportive, and translation of specific medical therapies has not been successful. Zebrafish larvae offer a unique platform for drug screening to rapidly identify neuroprotective compounds following ICH. We applied the Spectrum Collection library compounds to zebrafish larvae acutely after ICH to screen for decreased brain cell death and identified 150 successful drugs. Candidates were then evaluated for possible indications with other cardiovascular diseases. Six compounds were identified, including two angiotensin-converting enzyme inhibitors (ACE-Is). Ramipril and quinapril were further assessed to confirm a significant 55% reduction in brain cell death. Proteomic analysis revealed potential mechanisms of neuroprotection. Using the INTERACT2 clinical trial dataset, we demonstrated a significant reduction in the adjusted odds of an unfavourable shift in the modified Rankin scale at 90 days for patients receiving an ACE-I after ICH (versus no ACE-I; odds ratio, 0.80; 95% confidence interval, 0.68-0.95; P=0.009). The zebrafish larval model of spontaneous ICH can be used as a reliable drug screening platform and has identified therapeutics that may offer neuroprotection. This article has an associated First Person interview with the first author of the paper.

Keywords: ACE inhibitors; Drug screen; Intracerebral haemorrhage; Stroke; Zebrafish.

Fig. 1.

The drug screen resulted in 150 compounds that prevented brain cell death in a model of ICH. (A) Schematic of the screening protocol. (B) Structural similarities, based on MACCS keys, between the 150 positive hits represented by a radial tree. The six highlighted compounds were identified as predicted to be efficacious in cardiovascular disease. hpi, h post-injury; ICH, intracerebral haemorrhage; MACCS, Molecular Access System. Scale bars: 100 μm.

Fig. 2.

Angiotensin-converting enzyme inhibitor (ACE-I) treatment offers cellular neuroprotection but does not translate to functionality in zebrafish. (A) Representative images of annexinV expression in the brain at 72 h post-fertilisation (hpf) in DMSO-treated ICH− and ICH+ controls compared to ICH+ individuals that have been treated with ACE-I (n=10). (B) annexinV expression in the brain 24 h after ICH is significantly increased in DMSO vehicle-treated larvae, as previously observed in this model (***P=0.0003, ****P<0.0001) (Crilly et al., 2018; Crilly et al., 2019). Data (from n=10 larvae from three technical repeats) were analysed using a two-way ANOVA and Tukey's post hoc analysis. (C) Representative swimming traces for (n=3/24) larvae from each group, tracked over 10 min for spontaneous movement in response to a white-light stimulus every 1 min. Data presented from two technical repeats (clutches). (D,E) Functional analysis of spontaneous locomotion 72 h after ICH (120 hpf) shows an increase in time spent moving (D) and distance moved (E) in ramipril- and quinapril hydrochloride-treated ICH+ groups compared to DMSO vehicle-treated ICH+ groups, although these differences were not significant (P-values shown on graph). Data were analysed using mixed linear modelling and chi-squared test for non-parametric data. No significant difference between DMSO ICH+ and treatment ICH+ groups was observed (P=0.6); however, the data trend suggests that ICH+ treatment group values are closer to the ICH− treatment group values, highlighted by the increase in P-values presented on the graph. Scale bars: 100 μm.

Fig. 3.

Proteomic analysis highlights several pathways associated with neuroprotection from ACE-I treatment. (A) Significant changes in protein expression levels in the heads (n=5) of ramipril (left, green)- and quinapril (right, orange)-treated larvae when compared to DMSO controls. (B,C) Functional annotation performed using Database for Annotation, Visualization and Integrated Discovery (DAVID) identified gene ontology, protein pathway and functional category similarities, and the top ten upregulated pathways (top) and downregulated pathways (bottom), compared to total background detection for ramipril (B) and quinapril (C). Classification stringency set to ‘medium’, similarity threshold=0.5. Bars represent enrichment score and points (•) are the number of proteins detected in that pathway. P-values are presented.

Fig. 4.

Flowchart showing exposure to ACE-I at onset, day 1 and day 7 in patients participating in the INTERACT2 trial. Exposure to ACE-I is shown separately for those alive and dead at day 1 and day 7.

Fig. 5.

ACE-I given after ICH is associated with improved modified Rankin scale (mRS) scores at 90 days. Exposure to an ACE-I after ICH only, and within 7 days of ICH (bottom) [versus no ACE-I exposure (middle)], was associated with a favourable shift in functional outcomes in a multifactorial ordinal shift analysis (P=0.009). Exposure to an ACE-I at ICH onset (top, versus no ACE-I exposure) was not independently associated with a shift in the mRS score at 90 days (P=0.169).